Power grid section data retrieval method considering manifold sorting algorithm

A manifold sorting and data retrieval technology, applied in digital data information retrieval, electrical digital data processing, special data processing applications, etc., to avoid the problem of dimension disaster, improve accuracy, and improve similarity measurement.

Pending Publication Date: 2021-01-05
0 Cites 0 Cited by

AI-Extracted Technical Summary

Problems solved by technology

[0005] The main problem to be solved by the present invention is: Aiming at the problem that the efficiency of multi-dimensional query is not high during data retrieval, and the retrieval results cannot be matched in multiple dimensions as a whole, the present invention proposes a grid cross-section data retrieval method based on manifold sorting, using low-dimensional Data retrieval is carried out in the manifold subspace, the power grid secti...
View more

Method used

Embodiment 1, a kind of grid cross-section data query and retrieval method based on manifold sorting, the method is made up of five parts: utilize low-dimensional manifold subspace to carry out data retrieval, grid cross-section data is described as multi-dimensional vector space Corresponding points in , create a weighted graph model, improve the original similarity measure based on Euclidean distance, use belief propagation to assign ranking scores, and improve the accuracy of retrieval results. The present invention will be further described below using the accompanying drawings and examples.
In the experiment, this algorithm is compared with the original Manifold Ranking method, keyword retrieval and the retrieval method based on fuzzy rough set theory, and the Precision-Recall curve drawn. It can be seen from the figure that in terms ...
View more


Aiming at the problems of low multi-dimensional query efficiency and incapability of multi-dimensional integral matching of a retrieval result during data retrieval, the invention discloses a power grid section data retrieval method considering a manifold sorting algorithm. The method comprises the following steps: describing power grid section data into corresponding points in a multi-dimensionalvector space, and creating a weighted graph model; acquiring a retrieval result by considering the overall approximate manifold structure of the data, so the retrieval result has relatively high correlation with source query; and distributing the sorting scores by confidence propagation, so that the accuracy of a retrieval result is improved, and the defects of correlation measurement on high-dimensional data query processing are effectively avoided.

Application Domain

Digital data information retrievalCharacter and pattern recognition +1

Technology Topic

Data querySorting algorithm +9


  • Power grid section data retrieval method considering manifold sorting algorithm
  • Power grid section data retrieval method considering manifold sorting algorithm
  • Power grid section data retrieval method considering manifold sorting algorithm


  • Experimental program(1)
  • Effect test(1)

Example Embodiment

[0026] Embodiment 1, a method for querying and retrieving power grid cross-section data based on manifold sorting, the method consists of five parts: using low-dimensional manifold subspace for data retrieval, and describing power grid cross-section data as corresponding data in a multi-dimensional vector space. Points, create a weighted graph model, improve the similarity measure based on Euclidean distance, use belief propagation to assign ranking scores, and improve the accuracy of retrieval results. The present invention will be further described below with reference to the accompanying drawings and embodiments.
[0027] (1) Describe the grid cross-section data as corresponding points in a multi-dimensional vector space
[0028] Map the power data in the dataset to corresponding points in the vector space and create a K-NN graph.
[0029] (2) Create a weighted graph model
[0030] Compute node x in K-NN graph i and x j The weight W of the edge between ij , if there is no edge, then W ij =0, so as to obtain the weight matrix; normalize the weight matrix to obtain the similarity matrix S=D -1/2 WD -1/2 , where D is a diagonal matrix, and normalizing W can make the propagation rules converge.
[0031] (3) Iterative calculation
[0032] f(t+1)=αSf(t)+(1-α)y (4)
[0033] where α∈[0,1), the size of the α value represents the proportion of the score contribution from adjacent nodes.
[0034] When the confidence of each node in the set is not updated, the iteration is completed, and the data corresponding to the first n nodes is returned to the user according to the final converged similarity value.
[0035] (4) Using the SCADA system monitoring data of the wind turbine as the experimental data set, each piece of data constitutes a cross section which includes multiple attributes such as average wind speed, average active power, maximum total power generation, and average ambient temperature. The data retrieval effect is measured by recall rate, precision rate and NDCG (Normalized Discount Cumulative Gain) indicators. Among them, the precision rate P represents the proportion of data actually related to the query in the first n pieces of data retrieved by a certain rank method, such as formula (5), where TP is the positive sample predicted by the model to be positive, and FP is the model predicted to be positive and negative. Sample; the recall rate R represents the degree to which the retrieval is thorough and expresses the coverage of the answer set to all the answers. The calculation formula is formula (6), where FN is the positive sample predicted by the model to be negative. In the experiment, the retrieval results of different algorithms are compared by drawing Precision-Recall curves. If a curve is above another curve, the method corresponding to the upper curve is better.
[0038] In this algorithm, four parameter values ​​need to be determined. The smoothing parameter α is used to control the contribution from the prior score and the score from the neighbor nodes to the final ranking score. The larger the value of α, the greater the contribution from the adjacent node score. The larger the proportion is; the heat kernel parameter σ; the number of neighbors K for constructing the weighted graph model, and the parameter value settings are shown in Table 1:
[0039] Table 1 Parameter settings


no PUM

Description & Claims & Application Information

We can also present the details of the Description, Claims and Application information to help users get a comprehensive understanding of the technical details of the patent, such as background art, summary of invention, brief description of drawings, description of embodiments, and other original content. On the other hand, users can also determine the specific scope of protection of the technology through the list of claims; as well as understand the changes in the life cycle of the technology with the presentation of the patent timeline. Login to view more.

Similar technology patents

Adaptive fault detection method for airplane rotation actuator driving device based on deep learning

InactiveCN104914851Aimprove accuracyReduce the false alarm rate of detection

Video monitoring method and system


Scene semantic segmentation method based on full convolution and long and short term memory units

InactiveCN107480726Aimprove accuracylow resolution accuracy

Classification and recommendation of technical efficacy words

  • improve accuracy

Golf club head with adjustable vibration-absorbing capacity

InactiveUS20050277485A1improve grip comfortimprove accuracy

Stent delivery system with securement and deployment accuracy

ActiveUS7473271B2improve accuracyreduces occurrence and/or severity

Method for improving an HS-DSCH transport format allocation

InactiveUS20060089104A1improve accuracyincrease benefit

Catheter systems

ActiveUS20120059255A1increase selectivityimprove accuracy

Gaming Machine And Gaming System Using Chips

ActiveUS20090075725A1improve accuracy
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products